Connecting weighted automata, tensor networks and recurrent neural networks through spectral learning

被引:1
|
作者
Li, Tianyu [1 ]
Precup, Doina [2 ]
Rabusseau, Guillaume [3 ,4 ]
机构
[1] McGill Univ, Mila, Montreal, PQ, Canada
[2] McGill Univ, CCAI Chair Mila, Montreal, PQ, Canada
[3] Univ Montreal, CCAI Chair Mila, Montreal, PQ, Canada
[4] Univ Montreal, DIRO, Montreal, PQ, Canada
关键词
Weighted automata; Spectral learning; Recurrent neural networks; Tensor networks; Tensor train decomposition; FINITE-STATE AUTOMATA; MATRIX RECOVERY;
D O I
10.1007/s10994-022-06164-1
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
In this paper, we present connections between three models used in different research fields: weighted finite automata (WFA) from formal languages and linguistics, recurrent neural networks used in machine learning, and tensor networks which encompasses a set of optimization techniques for high-order tensors used in quantum physics and numerical analysis. We first present an intrinsic relation between WFA and the tensor train decomposition, a particular form of tensor network. This relation allows us to exhibit a novel low rank structure of the Hankel matrix of a function computed by a WFA and to design an efficient spectral learning algorithm leveraging this structure to scale the algorithm up to very large Hankel matrices. We then unravel a fundamental connection between WFA and second-order recurrent neural networks (2-RNN): in the case of sequences of discrete symbols, WFA and 2-RNN with linear activation functions are expressively equivalent. Leveraging this equivalence result combined with the classical spectral learning algorithm for weighted automata, we introduce the first provable learning algorithm for linear 2-RNN defined over sequences of continuous input vectors. This algorithm relies on estimating low rank sub-blocks of the Hankel tensor, from which the parameters of a linear 2-RNN can be provably recovered. The performances of the proposed learning algorithm are assessed in a simulation study on both synthetic and real-world data.
引用
收藏
页码:2619 / 2653
页数:35
相关论文
共 50 条
  • [1] Connecting weighted automata, tensor networks and recurrent neural networks through spectral learning
    Tianyu Li
    Doina Precup
    Guillaume Rabusseau
    Machine Learning, 2024, 113 : 2619 - 2653
  • [2] Connecting Weighted Automata and Recurrent Neural Networks through Spectral Learning
    Rabusseau, Guillaume
    Li, Tianyu
    Precup, Doina
    22ND INTERNATIONAL CONFERENCE ON ARTIFICIAL INTELLIGENCE AND STATISTICS, VOL 89, 2019, 89
  • [3] Distillation of weighted automata from recurrent neural networks using a spectral approach
    Eyraud, Remi
    Ayache, Stephane
    MACHINE LEARNING, 2024, 113 (05) : 3233 - 3266
  • [4] Distillation of weighted automata from recurrent neural networks using a spectral approach
    Rémi Eyraud
    Stéphane Ayache
    Machine Learning, 2024, 113 : 3233 - 3266
  • [5] Evaluating the Learning of Automata through the Use of Recurrent Neural Networks
    Lima, L.
    Sampaio, A.
    IEEE LATIN AMERICA TRANSACTIONS, 2018, 16 (10) : 2609 - 2616
  • [6] Learning minimal automata with recurrent neural networks
    Aichernig, Bernhard K.
    Koenig, Sandra
    Mateis, Cristinel
    Pferscher, Andrea
    Tappler, Martin
    SOFTWARE AND SYSTEMS MODELING, 2024, 23 (03): : 625 - 655
  • [7] Recurrent neural networks and finite automata
    Siegelmann, HT
    COMPUTATIONAL INTELLIGENCE, 1996, 12 (04) : 567 - 574
  • [8] Constrained Training of Recurrent Neural Networks for Automata Learning
    Aichernig, Bernhard K.
    Koenig, Sandra
    Mateis, Cristinel
    Pferscher, Andrea
    Schmidt, Dominik
    Tappler, Martin
    SOFTWARE ENGINEERING AND FORMAL METHODS, SEFM 2022, 2022, 13550 : 155 - 172
  • [9] Weighted automata extraction and explanation of recurrent neural networks for natural language tasks
    Wei, Zeming
    Zhang, Xiyue
    Zhang, Yihao
    Sun, Meng
    JOURNAL OF LOGICAL AND ALGEBRAIC METHODS IN PROGRAMMING, 2024, 136
  • [10] AdaAX: Explaining Recurrent Neural Networks by Learning Automata with Adaptive States
    Hong, Dat
    Segre, Alberto Maria
    Wang, Tong
    PROCEEDINGS OF THE 28TH ACM SIGKDD CONFERENCE ON KNOWLEDGE DISCOVERY AND DATA MINING, KDD 2022, 2022, : 574 - 584